Over the past year, I’ve started delegating more of my personal and work tasks to ChatGPT. If I find myself unable to delegate something due to poor reception or another silly reason, a small wave of irritation stirs inside me. Convenience is the path of least resistance, but what if resistance is sometimes necessary?
Calculators have made it harder for us to do mental math. Finding your way without a GPS navigator has become a relic of the past, and the Internet has generally made us worse at remembering information because a) it’s always available, and b) its availability has caused a surge in supply.
Sure, all these things have brought more benefits than harm, but the trade-offs we made can’t be ignored. There’s also a risk that the past is no predictor of the future, especially when it comes to AI systems.
| Invention | Impact on Cognitive Abilities |
|---|---|
| Calculators | Reduced ability for mental calculations |
| GPS Navigators | Decreased sense of spatial navigation |
| Internet | Diminished memory retention |
| Artificial Intelligence | Potential trade-offs in analytical and problem-solving abilities |
And I’m not talking about AI enslaving or wiping us off the face of the planet to establish the existence of AGI. Although the above can be the AGI’s goal, my intention is to emphasize the trade-offs. They should be made based on the completeness of information, making the unconscious conscious.
For instance, if AI boosts global productivity but costs us a few IQ points, is that a reasonable trade-off? Personally, I wouldn’t want to make that bet, as we don’t know what impact generative systems will have on our perception and analytical abilities. No wonder OpenAI is actively hiring Economic Impact Researchers, which suggests they don’t know either but are very keen to find out. Reflecting on this made me consider how we might gain productivity without sacrificing cognitive abilities.
Work vs. “the” work
One crucial aspect to explore is the difference between work that only matters for its result and work where the process is equally, if not more, important. The process is crucial because the task forces you to engage your brain, creating new neural connections and making you a better thinker or problem-solver.
The problem is that the line between these types of work is becoming increasingly blurred. It becomes even more complex due to the subjective perception of a task’s importance. A task that I consider a necessary mental exercise might be perceived as just a “check-the-box” item by someone in a different field or profession. For example, I consider user data analysis a necessary mental exercise for every product person to uncover behavioral user patterns necessary for improving the product. However, a small business owner may rely solely on automated reports, as data analysis is more of a routine task for them.
Obviously, children will be most affected by this blurred distinction, as they still lack the ability to understand the relative importance of different assignments. Dopamine pathways will do what they always do best: make us addicted to a certain type of behavior, and then it’s really difficult to reverse the flywheel effect. But who am I kidding? We, adults, sometimes are worse than children, so the risk is incredibly high across all groups.
It’s easy to fall into the trap of feeling unprecedented productivity from having AI do the work for us. But this work might be necessary to prepare you for more complex tasks in the future. No one will teach us multiplication and division if we haven’t first figured out addition and subtraction. A more productive world does not necessarily mean a more creative or original one, and that’s worth remembering.
Making yourself think vs. making AI think for you
There’s no shortage of problems in the world to tackle, so everything ultimately comes down to prioritization. In management, there’s a fairly simple 2×2 delegation matrix where individual readiness is on the vertical axis, and task importance is on the horizontal axis.
So I’ve decided that if my task is unimportant or less important and the AI’s readiness to solve it is high—i.e., ChatGPT can handle it through probability distribution or because it has seen similar information during its training—I delegate it to AI. If the task is important and the system’s readiness is low—usually tasks that require originality or creativity—I do it myself but may use AI as a coach or mentor. A few prompts that I shared below can help with this.
This approach isn’t perfect, as it still relies on a subjective assessment of a task’s importance, but it’s definitely better than nothing. The fact that different people will prioritize the same task differently is a feature of the system, not a bug. Adam Smith once discovered that specialization is the engine of economic progress. By dividing the labor, the productivity of the whole system increase. So, diversity of perception should be an advantage, not a drawback.
The greatest improvements in the productive powers of labour, and the greater part of the skill, dexterity, and judgment with which it is anywhere directed or applied, seem to have been the effects of the division of labour.
The Wealth of Nations (1776)
Therefore, I hope we will make AI push us to think more, not less. And through more thinking, we will create new, original approaches or solutions to old but very important problems.
Bonus: 10 thought-provoking ChatGPT prompts
1. I want to write [format] about [topic]. Ask me questions one by one to help me explain my idea clearly.
2. Rewrite this in the style of [favorite writer/author], adding details to match their way of writing.
3. Find long words or tricky phrases in this text. Suggest shorter, easier words or phrases.
4. Give me a tough critique of my [ideas/writing/code]. Be very honest and don't hold back.
5. List all the doubts or criticisms people might have about this [writing/idea]. Suggest ways to answer them.
6. Look at the logical flow of my argument in [specific work]. Point out any mistakes or missing parts.
7. Look at [specific situation/idea] and find any blind spots or missed details. Tell me what might be missing.
8. Look at [specific problem] from another field's point of view (like sociology, economics, psychology). What new ideas or solutions do you see?
9. Find and critique any logical mistakes or biases in my argument for [specific idea]. Suggest ways to make the argument stronger.
10. Create a detailed argument against [popular belief/assumption]. Use evidence and logic to support your counter-argument.